T cells recognize a glycopeptide derived from type II collagen in a model for rheumatoid arthritis

Article abstract of DOI:10.1021/ja980489k

Even though most eucaryotic proteins are glycosylated, very little is known on if, or how, the glycans influence essential immunological events such as antigen processing, major histocompatibility complex (MHC) restricted presentation, and recognition by T cells. We have used synthetic glycopeptides to elucidate the specificity of T cell hybridomas, obtained by immunization with the glycoprotein type II collagen in a mouse model for rheumatoid arthritis. To enable these studies, glycosylated and suitably protected derivatives of (5R)-5-hydroxy-L-lysine, and the similar 5-hydroxy- L-norvaline, were prepared and then used in Fmoc solid-phase synthesis of glycopeptides related to the immunodominant fragment from type II collagen, CII(256-270). Evaluation of the synthetic glycopeptides provided evidence that antigen-presenting cells can indeed process glycoproteins to glycopeptides, which elicit a T cell response when presented by class II MHC molecules. A glycopeptide carrying a single β-D-galactosyl residue attached to hydroxylysine at position 264 in the center of the CII(256-270) peptide was recognized by most of the hybridomas in a way involving specific contacts between the carbohydrate and the T cell receptor. The results suggest an explanation for the recent observation that glycosylated type II collagen induces more severe forms of arthritis in the mouse than deglycosylated type II collagen and provide additional knowledge on how rheumatoid arthritis may occur also in humans.

Full text of DOI:10.1021/ja980489k

7676
J. Am. Chem. Soc. 1998, 120, 7676-7683
T Cells Recognize a Glycopeptide Derived from Type II Collagen in
a Model for Rheumatoid Arthritis
Johan Broddefalk,† Johan Ba1cklund,‡ Fredrik Almqvist,† Martin Johansson,§
Rikard Holmdahl,‡ and Jan Kihlberg*,†
Contribution from Organic Chemistry, Umeå UniVersity, S-901 87 Umeå, Sweden, the Department of Cell
& Molecular Biology, Section for Medical Inflammation Research, Lund UniVersity, P.O. Box 94,
S-221 00 Lund, Sweden, and Organic Chemistry 2, Center for Chemistry and Chemical Engineering,
Lund Institute of Technology, Lund UniVersity, P.O. Box 124, S-221 00 Lund, Sweden
ReceiVed February 13, 1998
Abstract: Even though most eucaryotic proteins are glycosylated, very little is known on if, or how, the glycans
influence essential immunological events such as antigen processing, major histocompatibility complex (MHC)
restricted presentation, and recognition by T cells. We have used synthetic glycopeptides to elucidate the
specificity of T cell hybridomas, obtained by immunization with the glycoprotein type II collagen in a mouse
model for rheumatoid arthritis. To enable these studies, glycosylated and suitably protected derivatives of
(5R)-5-hydroxy-L-lysine, and the similar 5-hydroxy-L-norvaline, were prepared and then used in Fmoc solid-
phase synthesis of glycopeptides related to the immunodominant fragment from type II collagen, CII(256-
270). Evaluation of the synthetic glycopeptides provided evidence that antigen-presenting cells can indeed
process glycoproteins to glycopeptides, which elicit a T cell response when presented by class II MHC molecules.
A glycopeptide carrying a single â-D-galactosyl residue attached to hydroxylysine at position 264 in the center
of the CII(256-270) peptide was recognized by most of the hybridomas in a way involving specific contacts
between the carbohydrate and the T cell receptor. The results suggest an explanation for the recent observation
that glycosylated type II collagen induces more severe forms of arthritis in the mouse than deglycosylated
type II collagen and provide additional knowledge on how rheumatoid arthritis may occur also in humans.
Introduction
events that all are essential in elimination of the infection.
Peptides derived by processing of endogenous proteins are also
presented to T cells, but various mechanisms normally ensure
that they do not elicit an immune response, thereby preventing
occurrence of autoimmune disease.
Although most eucaryotic proteins and many viral proteins
are glycosylated, very little is known concerning if, or how,
the glycans influence antigen processing, MHC-restricted
presentation, and T cell recognition.2 In fact, prior to the
beginning of the 1990s carbohydrates were regarded only as T
cell independent antigens; i.e., it was assumed that T cells were
unable to recognize carbohydrates. Recent model studies with
synthetic neoglycopeptides have, however, revealed that mono-
or small oligosaccharides can be attached to T cell immunogenic
peptides without loss of MHC binding, provided that the position
for the glycan is chosen carefully.3 These studies have also
Processing of protein antigens into short peptides by antigen-
presenting cells is critical for proper function of the immune
system of higher vertebrates.1 In higher organisms most cells
are able to process foreign or altered protein antigens produced
intracellularily, for instance due to viral infections or malignant
transformations. Peptides resulting from this degradation are
bound by class I major histocompatibility complex (MHC)
molecules, and the complexes are then transported to the cell
surface where they are displayed to circulating T cells.
Recognition of peptide-class I MHC complexes by receptors
on “cytotoxic” (CD8+) T cells triggers release of compounds
which selectively kill the presenting cell, thereby eliminating
the viral infection or the malignant transformation. In contrast,
specialized antigen-presenting cells, such as macrophages, take
up and process extracellular protein material, e.g., bacterial
proteins. Recognition of complexes between peptides and class
II MHC molecules on the surface of such antigen-presenting
cells by “helper” (CD4+) T cells elicits release of immuno-
modulating cytokines, such as interleukines. These cytokines
are essential for inducing isotype switching from production of
IgM to high-affinity IgG antibodies in B cells, as well as for
production of memory B cells and activation of phagocytic cells;
(2) Reviewed in (a) Kihlberg, J.; Elofsson, M. Curr. Med. Chem. 1997,
4, 79-110. (b) Carbone, F. R.; Gleeson, P. A. Glycobiology 1997, 7, 725-
730.
(3) (a) Ishioka, G. Y.; Lamont, A. G.; Thomson, D.; Bulbow, N.; Gaeta,
F. C. A.; Sette, A.; Grey, H. M. J. Immunol. 1992, 148, 2446-2451. (b)
Haurum, J. S.; Arsequell, G.; Lellouch, A. C.; Wong, S. Y. C.; Dwek, R.
A.; McMichael, A. J.; Elliot, T. J. Exp. Med. 1994, 180, 739-744. (c) Otvos
Jr., L.; Urge, L.; Xiang, Z. Q.; Krivulka, G. R.; Nagy, L.; Szendrei, G. I.;
Ertl, H. C. J. Biochim. Biophys. Acta 1994, 1224, 68-76. (d) Deck, B.;
Elofsson, M.; Kihlberg, J.; Unanue, E. R. J. Immunol. 1995, 155, 1074-
1078. (e) Abdel-Motal, U. M.; Berg, L.; Rose´n, A.; Bengtsson, M.; Thorpe,
C. J.; Kihlberg, J.; Dahme´n, J.; Magnusson, G.; Karlsson, K.-A.; Jondal,
M. Eur. J. Immunol. 1996, 26, 544-551. (f) Jensen, T.; Galli-Stampino,
L.; Mouritsen, S.; Frische, K.; Peters, S.; Meldal, M.; Werdelin, O. Eur. J.
Immunol. 1996, 26, 1342-1349. (g) Jensen, T.; Hansen, P.; Galli-Stampino,
L.; Mouritsen, S.; Frische, K.; Meinjohanns, E.; Meldal, M.; Werdelin, O.
J. Immunol. 1997, 158, 3769-3778.
† Umeå University.
‡ Cell & Molecular Biology, Lund University.
§ Organic Chemistry 2, Lund University.
(1) (a) Grey, H. M.; Sette, A.; Buus, S. Sci. Am. 1989, NoV, 38-46. (b)
Engelhard, V. H. Sci. Am. 1994, Aug, 44-51. (c) Monaco, J. J. J. Leukocyte
Biol. 1995, 57, 543-547. (d) Janeway, C. A., Jr.; Travers, P. Immunobi-
ology: the immune system in health and disease, 2nd ed.; Current Biology
Ltd./Garland Publishing Inc.: London/New York, 1996.
S0002-7863(98)00489-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/23/1998

Glycopeptide DeriVed from Type II Collagen
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7677
demonstrated that immunization of mice with neoglycopeptides
can elicit T cells which specifically recognize the carbohydrate
moiety if this is located in the center of the peptide. For
glycoproteins, such as HIV gp120 and influenza virus hemag-
glutinin which carry large N-linked carbohydrate moieties,
indirect evidence has been obtained suggesting that the glycans
influence uptake in antigen-presenting cells, as well as process-
ing and MHC restricted presentation of peptide epitopes derived
from the glycoproteins.4 In addition, desialylated ovine sub-
maxillary mucin (A-OSM), which carries multiple copies of the
O-linked tumor-associated Tn antigen (GalNAcR-Ser/Thr), was
found to elicit a carbohydrate dependent cellular immune
response.5 However, as for the N-linked glycoproteins, the role
of the carbohydrate in influencing uptake in or processing by
antigen-presenting cells, or as part of the antigenic determinant
being recognized by the T cell receptor, was not established.
We now present a study of the response and carbohydrate
specificity of helper T cell hybridomas obtained in mice after
immunization with the glycoprotein type II collagen. Im-
munization of mice with type II collagen induces an inflam-
matory response, termed collagen-induced arthritis (CIA).6 CIA
is accompanied by erythema and severe, painful swelling of
peripheral joints, i.e., with symptoms and histopathology similar
to those displayed by patients suffering from rheumatoid
arthritis. It should be emphasized that type II collagen is found
mainly in joint cartilage, in contrast to type I collagen which is
the major constituent in connective tissue, and that CIA is the
most frequently used model for rheumatoid arthritis. In mice
the immunization leads to a type II collagen specific and class
II MHC (Aq) restricted T cell response. Since rheumatoid
arthritis in humans is correlated to the presence of certain class
II MHC molecules, it is possible that a T cell response to type
II collagen plays a role also in humans for development of this
autoimmune disease. We previously showed that a panel of T
cell hybridomas,7 obtained from mice immunized with type II
collagen, recognized an immunodominant T cell epitope located
within a polypeptide fragment of type II collagen which
contained residues 256-270 [CII(256-270)].7a Surprisingly,
even though all of the T cell hybridomas recognized type II
collagen and proteolytic fragments containing CII(256-270)
after antigen processing, only a few of them responded to the
synthetic peptide CII(256-270) (cf. 12, Figure 1) or to related
synthetic peptides.7b This observation could be due to the fact
that lysine residues in collagen, e.g., those at positions 264 and
270 of CII(256-270), can undergo posttranslational hydrox-
ylation and subsequent glycosylation by â-D-galactopyranosyl
or R-D-glucopyranosyl-(1-2)-â-D-galactopyranosyl moieties.8
The importance of posttranslational modifications was suggested
Figure 1. Peptides 12 and 13 and glycopeptides 14-17, which
correspond to residues 256-270 of type II collagen, were used to
delineate the response and specificity of T cell hybridomas obtained
from mice immunized with this O-linked glycoprotein.
by the observation that chemical removal of carbohydrates from
type II collagen resulted in loss of recognition by most of the
T cell hybridomas which did not respond to the synthetic CII-
(256-270) peptide.7b In addition, deglycosylated type II
collagen was found to induce less severe arthritis in the mouse
than the native form of the glycoprotein. However, as in
previous studies of the cellular immune response elicited by
glycoproteins,4,5 it was not established if the carbohydrate
moieties influenced antigen processing, MHC restricted pre-
sentation, T cell recognition, or other factors.
Results and Discussion
To delineate the role(s) of the posttranslational hydroxylation
and glycosylation of lysine in type II collagen for the T cell
immune response elicited by this O-linked glycoprotein, we have
prepared protected (5R)-5-hydroxy-L-lysine 3 and the corre-
sponding â-galactoside 8, as well as the glycosylated derivatives
6 and 11 of the structurally related amino acid 5-hydroxy-L-
norvaline (cf. Scheme 1). These building blocks were then used
for synthesis of peptides and glycopeptides related to type II
collagen (cf. 12-17, Figure 1). (5R)-5-Hydroxy-L-lysine was
protected by transformation into a cupric chelate which allowed
regioselective tBoc protection of the ꢀ-amino group.9 After
dissociation of the chelate the R-amino and carboxyl groups
were protected with Fmoc and benzyl groups,10 respectively,
to give 2. Silylation of the hydroxyl group of 2 followed by
selective11 cleavage of the benzyl ester gave hydroxylysine
building block 3, which was used for synthesis of peptide 13
having a hydroxylysine residue at position 264 of type II
collagen. As reported recently by us, â-galactosylation of
protected 5-hydroxy-L-norvaline 1 failed using galactosyl donors
having participating groups at O-2.12 However, epoxidation of
(4) (a) Doe, B.; Steimer, K. S.; Walker, C. M. Eur. J. Immunol. 1994,
24, 2369-2376. (b) Botarelli, P.; Houlden, B. A.; Haigwood, N. L.; Servis,
C.; Montagna, D.; Abrignani, S. J. Immunol. 1991, 147, 3128-3132. (c)
Barnett, B. C.; Graham, C. M.; Burt, D. S.; Skehel, J. J.; Thomas, D. B.
Eur. J. Immunol. 1989, 19, 515-521. (d) Thomas, D. B.; Hodgson, J.; Riska,
P. F.; Graham, C. M. J. Immunol. 1990, 144, 2789-2794. (e) Drummer,
H. D.; Jackson, D. C.; Brown, L. E. Virology 1993, 192, 282-289. (f)
Brown, L. E.; White, D. O.; Jackson, D. C. J. Virol. 1993, 67, 2887-2893.
(g) Jackson, D. C.; Drummer, H. E.; Urge, L.; Otvos Jr., L.; Brown, L. E.
Virology 1994, 199, 422-430.
(5) Singhal, A.; Fohn, M.; Hakomori, S.-i. Cancer Res. 1991, 51, 1406-
1411.
(6) (a) Trentham, D. E.; Townes, A. S.; Kang, A. H. J. Exp. Med. 1977,
146, 857-868. (b) Holmdahl, R.; Andersson, M.; Goldschmidt, T. J.;
Gustafsson, K.; Jansson, L.; Mo, J. A. Immunol. ReV. 1990, 118, 193-
232.
(9) Scott, J. W.; Parker, D.; Parrish, D. R. Synth. Commun. 1981, 11,
303-314.
(10) Wang, S.-S.; Gisin, B. F.; Winter, D. P.; Makofske, R.; Kulesha, I.
D.; Tzougraki, C.; Meienhofer, J. J. Org. Chem. 1977, 42, 1286-1290.
(11) (a) Carpino, L. A.; Tunga, A. J. Org. Chem. 1986, 51, 1930-1932.
(b) Cleavage of the Fmoc group may occur during hydrogenolysis, but use
of ethyl acetate as solvent and adjustment of the amount of Pd/C allowed
selective removal of the benzyl ester to give 3. Selective debenzylation
was also possible in preparation of the glycosylated amino acids 6, 8, and
11.
(7) (a) Michae¨lsson, E.; Andersson, M.; Engstro¨m, Å.; Holmdahl, R. Eur.
J. Immunol. 1992, 22, 1819-1825. (b) Michae¨lsson, E.; Malmstro¨m, V.;
Reis, S.; Engstro¨m, Å.; Burkhardt, H.; Holmdahl, R. J. Exp. Med. 1994,
180, 745-749.
(12) (a) Broddefalk, J.; Bergquist, K.-E.; Kihlberg, J. Tetrahedron Lett.
1996, 37, 3011-3014. (b) Broddefalk, J.; Bergquist, K.-E.; Kihlberg, J.
Tetrahedron, submitted for publication.
(8) Spiro, R. G. J. Biol. Chem. 1967, 242, 4813-4823.

7678 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Broddefalk et al.
Scheme 1a
unsuccessful. Regioselective hydrogenolysis of the benzyl ester
in 10 was performed over Pd/C in the presence of ammonium
acetate,17 without simultaneous cleavage of the Fmoc or
p-methoxybenzyl groups. This gave building block 11, used
for synthesis of glycopeptide 16 which carries disaccharide
moieties both at position 264 and at position 270 of CII(256-
270).
Synthesis of peptide 13 and glycopeptide 15 was performed
on a polystyrene resin grafted with poly(ethylene glycol) spacers
(TentaGel resin) in an automatic peptide synthesizer. Synthesis
was performed according to the Fmoc strategy2a,18 under
conditions identical to those reported12 for synthesis of glyco-
peptides 14, 16, and 17, as well as other glycopeptides19
prepared recently by us. Use of acid labile protective groups
for the carbohydrate residues of the target glycopeptides had
the advantage that both the glycan and the peptide moieties were
deprotected simultaneously with acid-catalyzed cleavage from
the resin, allowing glycopeptides 14-17 to be isolated in 26-
45% yields after a single purification by reversed-phase HPLC.
The glycopeptides were characterized by 1H NMR spectroscopy,
fast atom bombardment mass spectroscopy, and amino acid
1
analysis (cf. H NMR data for glycopeptide 15 in Table 1).
Previously, dipeptides containing hydroxylysine have been
glycosylated and then used for solution synthesis of short
glycopeptides,20 but the present work respresents the first use
of a glycosylated building block of hydroxylysine in Fmoc solid-
phase synthesis.
Use of peptides 12 and 13 together with glycopeptides 14-
17 allowed a detailed investigation of the specificity of the MHC
class II restricted T cell hybridomas7 obtained after immuniza-
tion of mice with type II collagen. To this end each of the 29
hybridomas was incubated with spleen cells as antigen-present-
ing cells and dilution series of peptides 12 and 13, glycopeptides
14-17, as well as type II collagen as control. After 24 h the
response of each hybridoma, i.e., interleukin 2 (IL-2) secreted
into the supernatant as a result of recognition of a (glyco)-
peptide-MHC complex, was determined in a standard radio-
assay21 based on proliferation of the IL-2 sensitive T cell clone
CTLL. Seven of the hybridomas raised to type II collagen did
not respond to any of the glycopeptides, but six of these instead
recognized peptide 12 which has lysine at position 264, whereas
one responded only to peptide 13 having hydroxylysine at this
position (cf. hybridoma HDBR.1, Figure 2a). The response
of the remaining 22 hybridomas appeared to depend on
carbohydrate since they all recognized type II collagen but not
peptides 12 and 13. Interestingly, as many as 20 of the 22
hybridomas responded to glycopeptide 15, in which hydroxy-
lysine carrying a â-D-galactopyranosyl residue has been incor-
porated at position 264 of CII(256-270) (cf. hybridomas
a Reagents and conditions: (a) (i) CuCO3‚Cu(OH)2‚H2O, tBoc2O,
H2O-dioxane then Chelex 100 (H+-form), (ii) FmocCl, Na2CO3, H2O-
dioxane 1:1, (iii) Cs2CO3, benzyl bromide, DMF. (b) (i) TBDMSOTf,
2,6-lutidine, CH2Cl2, 0 °C, (ii) H2, 10% Pd/C, EtOAc. (c) (i)
dimethyldioxirane, acetone, CH2Cl2, 0 °C, (ii) ZnCl2, THF, AW-300,
-50 °C f room temp. (d) H2, 10% Pd/C, EtOAc. (e) NIS, AgOTf,
CH2Cl2, 4 Å MS, -45 f -15 °C. (f) H2, 10% Pd/C, NH4OAc, EtOAc.
galactal 413 (1.2 equiv) with dimethyldioxirane gave the
corresponding R-1,2-anhydrosugar,14 which under zinc chloride
promotion gave â-galactoside 5 (50%) on reaction with 1.
Galactosylated hydroxylysine 7 was prepared from 2 in the same
manner by glycosylation with 4 (37% yield).15 Regioselective11
removal of the benzyl ester in 5 and 7 then gave the glycosylated
amino acids 6 and 8. These were subsequently used for
synthesis of glycopeptides 14 and 15 which have a â-D-
galactopyranosyl residue at position 264 of CII(256-270).
Glucosyl donor 9 allowed a highly stereoselective glycosylation
of the hydroxynorvaline galactoside 5 when using N-iodosuc-
cinimide and silver triflate as promoters,16 and the R-glucoside
10 was obtained in a high yield (71%).12b Regrettably, all
attempts to attach an R-D-glucosyl moiety to the galactosylated
hydroxylysine 7 by this, or other procedures, have so far been
(17) Sajiki, H. Tetrahedron Lett. 1995, 36, 3465-3468.
(18) Reviewed in: (a) Kunz, H. Angew. Chem., Int. Ed. Engl. 1987, 26,
294-308. (b) Paulsen, H. Angew. Chem., Int. Ed. Engl. 1990, 29, 823-
839. (c) Garg, H. G.; von dem Bruch, K.; Kunz, H. AdV. Carbohydr. Chem.
Biochem. 1994, 50, 277-310. (d) Meldal, M. In Glycopeptide synthesis;
Lee, Y. C., Lee, R. T., Eds.; Academic Press: San Diego, 1994; pp 145-
198. (e) Meldal, M. Curr. Opin. Struct. Biol. 1994, 4, 710-718. (f) Norberg,
T.; Lu¨ning, B.; Tejbrant, J. Methods Enzymol. 1994, 247, 87-106. (g)
Nakahara, Y.; Iijima, H.; Ogawa, T. In Stereocontrolled approaches to
O-glycopeptide synthesis; Kova´c, P., Ed.; American Chemical Society:
Washington, DC, 1994; pp 249-266. (h) Schultz, M.; Kunz, H. In Chemical
and enzymatic synthesis of glycopeptides; Jolle`s, P., Jo¨rnvall, H., Eds.;
Birkha¨user Verlag: Basel, 1995; pp 201-228. (i) Kihlberg, J.; Elofsson,
M.; Salvador, L. A. Methods Enzymol. 1997, 289, 221-245. (j) Arsequell,
G.; Valencia, G. Tetrahedron: Asymmetry 1997, 8, 2839-2876.
(19) Elofsson, M.; Salvador, L. A.; Kihlberg, J. Tetrahedron 1997, 53,
369-390.
(13) Alonso, R. A.; Vite, G. D.; McDevitt, R. E.; Fraser-Reid, B. J. Org.
Chem. 1992, 57, 573-584.
(14) (a) Halcomb, R. L.; Danishefsky, S. J. J. Am. Chem. Soc. 1989,
111, 6661-6666. (b) Danishefsky, S. J.; Bilodeau, M. T. Angew. Chem.,
Int. Ed. Engl. 1996, 35, 1380-1419.
(15) In the synthesis of both 5 and 7, the excess of galactal 4 had to be
adjusted carefully so as to avoid galactosylation of the secondary hydroxyl
group in the products. The decreased reactivity of the secondary hydroxyl
group in hydroxylysine 2, as compared to the primary hydroxyl group of
hydroxynorvaline 1, resulted in formation of larger amounts of higher order
glycosides in the synthesis of 7, thus lowering the yield of 7.
(16) Konradsson, P.; Udodong, U. E.; Fraser-Reid, B. Tetrahedron Lett.
1990, 31, 4313-4316.
(20) Koeners, H. J.; Schattenkerk, C.; Verhoeven, J. J.; van Boom, J. H.
Tetrahedron 1981, 37, 1763-1771.
(21) Gillis, S.; Smith, K. A. Nature 1977, 268, 154-156.

Glycopeptide DeriVed from Type II Collagen
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7679
Table 1. 1H NMR Data (δ, ppm) for Glycopeptide 15 in Water Containing 10% D2Oa
residue
NH
H-R
H-â
H-γ
H-δ
others
Gly256
Glu257
Hyp258
Gly259
Ile260
3.78, 3.63
4.62
8.73
1.99, 1.79
2.30, 1.03
2.24b
4.59
4.48
3.81b
0.78
8.69
8.08
8.64
7.79
8.17
8.48
8.44
8.48
8.64
8.41
3.89b
4.14
1.83
1.33
1.38, 1.14
0.87 (â-CH3)
Ala261
Gly262
Phe263
Hyl264
Gly265
Glu266
Gln267
Gly268
Pro269
Lys270
4.24
3.84b
4.53
3.04b
1.96, 1.66
7.31, 7.20 (arom)
4.24
1.54b
3.97
3.12, 2.91 (Hꢀ), Galâc
3.84b
4.24
2.02, 1.87
2.11, 1.93
2.24b
2.33b
4.31
4.12, 3.92
4.37
7.61, 6.92 (CONH2)
2.21, 1.92
1.78, 1.65
1.97b
1.37b
3.55b
1.61b
8.15
4.10
2.93b(Hꢀ), 7.55 (ꢀ-NH2)
a Obtained at 500 MHz, 278 K, and pH 5.4 with H2O as internal standard (δH 4.98 ppm). b Degeneracy has been assumed. c Chemical shifts (δ,
ppm) for the galactose moiety: 4.38 (H-1), 3.86 (H-4), 3.71b (H-6), 3.66 (H-5), 3.61 (H-3), 3.46 (H-2).
Figure 2. Response of selected T cell hybridomas on incubation with antigen-presenting spleen cells and increasing concentrations of the antigens
type II collagen (CII), peptides 12 and 13, and glycopeptides 14-17. Recognition of (glyco)peptide-MHC complexes on the surface of an antigen-
presenting cell by a T cell hybridoma results in secretion of the cytokine interleukin 2 (IL-2) in a dose dependent manner. This T cell response is
subsequently determined in a radioassay based on proliferation of the IL-2 sensitive T cell clone CTLL.21 (a) Hybridoma HDBR.1 responds to
peptide 13 which has hydroxylysine at position 264 of CII(256-270) but not when lysine or â-D-galactosylated hydroxylysine is found at this
position (cf. 12 and 15). (b) Hybridoma HCQ.10 responds when â-D-galactosylated hydroxylysine is found at position 264 of CII(256-270) (cf.
15), but it does not respond to peptides 12 and 13 or to glycopeptides 14 and 17. This hybridoma represents 17 of the 22 hybridomas which require
CII to be glycosylated. (c) Hybridoma HM1R.1 requires a â-D-galactopyranosyl residue to be attached to either hydroxylysine or hydroxynorvaline
at position 264 (cf. 15 and 14). Two other hybridomas respond in the same manner. (d) Hybridoma HCQ.11 is specific for an R-D-glucopyranosyl-
â-D-galactopyranosyl residue attached to hydroxynorvaline at position 264 (cf. 16 and 17).
HCQ.10 and HM1R.1 (Figure 2b,c). These 20 hybridomas did
not respond to glycopeptides 16 and 17 which carry an R-D-
glucopyranosyl-(1-2)-â-D-galactopyranosyl residue attached to
hydroxynorvaline 264, but three of them did respond to
glycopeptide 14, which has a galactosyl residue linked to
hydroxynorvaline 264 (cf. hybridoma HM1R.1, Figure 2c).
Hybridoma HCQ.11, one of the two hybridomas which did not
respond to the galactosylated glycopeptides 14 and 15, was
instead stimulated by glycopeptides 16 and 17 having R-D-
glucopyranosyl-(1-2)-â-D-galactopyranosyl residues attached

7680 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Broddefalk et al.
of the carbohydrate dependent T cell hybridomas, in contrast
to the nonglycosylated peptides 12 and 13 (Figure 2b). The
fact that only 3 of the 20 hybridomas which responded to 15
also recognized glycopeptide 14 (â-D-galactosylated on hydroxy-
norvaline) reveals the additional importance of the ꢀ-amino
group of hydroxylysine in contacting the T cell receptor (Figure
2c). Specific recognition of the carbohydrate moiety is further
confirmed by the inability of these three hybridomas to respond
when an R-D-glucopyranosyl-(1-2)-â-D-galactopyranosyl resi-
due was attached instead of â-D-galactose to hydroxynorvaline
264 (glycopeptide 17). The equal response displayed by clone
HCQ.11 to glycopeptides 16 and 17 indicates that glycosylation
of hydroxylysine at position 270 has no, or limited, importance
for the T cell response (Figure 2d). Residue 270 is therefore
most likely outside the determinant recognized by the T cell
receptor.
It is interesting to note that the majority of the carbohydrate
dependent T cell hybridomas (20 out of 22) respond when
galactose is attached to hydroxylysine 264 of CII(256-270)
whereas, so far, only one hybridoma recognizing an R-D-
glucopyranosyl-(1-2)-â-D-galactopyranosyl moiety at this posi-
tion has been identified. This T cell selectivity for recognition
of a mono- over a disaccharide at position 264 may have
different explanations. The glycosylation pattern of type II
collagen is not known, but a galactosyl residue might predomi-
nate at hydroxylysine 264. Alternatively, enzymatic removal
of the glucosyl moiety to give a galactosylated hydroxylysine
may occur during processing in the antigen-presenting cell. If
saccharide processing does not occur, another explanation could
be that the T cell receptor is less willing to accommodate the
larger disaccharide when forming the ternary complex with the
glycopeptide and the class II MHC molecule. As a result T
cells recognizing galactosylated glycopeptides would be selected
even if hydroxylysine 264 in collagen predominantly carried a
disaccharide moiety. Future studies with a synthetic CII(256-
270) glycopeptide having a disaccharide at hydroxylysine 264
should provide answers to some of these questions.
In conclusion, glycosylated derivatives of (5R)-5-hydroxy-
L-lysine, and the related 5-hydroxy-L-norvaline, have been
synthesized and used for preparation of glycopeptide fragments
from type II collagen by the Fmoc strategy. The glycopeptides
were used for evaluation of a panel of helper T cell hybridomas
obtained previously in collagen-induced arthritis, a mouse model
for rheumatoid arthritis. In this model disease is elicited by
immunization with type II collagen, the major protein of joint
cartilage. It was thus, for the first time, established that
immunization with a naturally occurring glycoprotein can elicit
carbohydrate specific T cells. In the present study, a glyco-
peptide carrying a single â-D-galactosyl residue attached to
hydroxylysine was recognized by most of the T cell hybridomas
obtained toward type II collagen, in a way involving direct
contacts between the carbohydrate and the T cell receptor. The
fact that most of the hybridomas responded to a glycopeptide,
but not to the corresponding peptide, suggests an explanation
for the recent observation7b that glycosylated type II collagen
induces more severe forms of arthritis in the mouse than
deglycosylated collagen. Finally, the possibility that T cells
directed to glycopeptide fragments from joint cartilage partici-
pate in development of autoimmune rheumatoid arthritis in
humans should not be overlooked.
Figure 3. A schematic description of the interactions formed in the
complex of glycopeptide 15, which corresponds to residues 256-270
of type II collagen, with the class II MHC molecule, and the T cell
receptor. According to peptide binding studies and computer modeling
Ile260 and Phe263 anchor the glycopeptide in the P1 and P4 pockets of
the MHC molecule, respectively.22 Galactosylated hydroxylysine 264
is then located optimally in the center of the glycopeptide for specific
interactions with the T cell receptor. This agrees well with the present
finding that galactosylation of hydroxylysine 264 is critical for
stimulation of the majority of the T cell hybridomas obtained by
immunization of mice with type II collagen in a model for rheumatoid
arthritis.
to hydroxynorvaline at positions 264 and 270, or only at 264
(Figure 2d). It can be assumed that this hybridoma, as well as
the one remaining nonresponding hybridoma, would have
recognized a glycopeptide having R-D-glucopyranosyl-(1-2)-
â-D-galactopyranose at hydroxylysine 264 if this had been
available through synthesis.
By performing binding studies using alanine-substituted
derivatives of peptide 12 and glycopeptide 14, we recently found
that Ile260 and Phe263 are the most important residues for
anchoring CII(256-270) into the binding groove of the Aq
molecule, i.e., the class II MHC molecule important in eliciting
an immune response to type II collagen in mice.22 In addition
it was found that glycosylation did not influence binding of CII-
(256-270) to the MHC molecule. Modeling based on the
peptide binding studies and the crystal structure23 of the human
class II MHC protein HLA-DR1 placed the MHC anchors, Ile260
and Phe263, in the P1 and P4 pockets of the MHC molecule,
respectively.22 As a consequence, the glycosylated residue 264
is located in the center of the peptide MHC-complex recognized
by the T cell receptor (Figure 3). Studies with synthetic
neoglycopeptides have shown such a location to be ideal for
direct contacts between the carbohydrate moiety and the T cell
receptor.2,3 Moreover, the recent crystal structure between a class
I MHC molecule, a peptide, and a T cell receptor revealed that
the side chain of a residue in the center of the peptide formed
critical contacts with the highly variable CD3 loops of the T
cell receptor.24 The importance of residue 264 in CII(256-
270) for the response of most of the T cell hybridomas therefore
appears to originate from contacts formed by the carbohydrate
moiety, and the side chain amino group, of this residue and the
CD3 loops of the T cell receptor. Specificity for the carbohy-
drate moiety is revealed by the ability of glycopeptide 15,
galactosylated on hydroxylysine, to elicit a response from most
(22) Kjelle´n, P.; Brunsberg, U.; Broddefalk, J.; Hansen, B.; Vestberg,
M.; Ivarsson, I.; Engstro¨m, Å.; Svejgaard, A.; Kihlberg, J.; Fugger, L.;
Holmdahl, R. Eur. J. Immunol. 1998, 28, 755-767.
Experimental Section
(23) Stern, L. J.; Brown, J. H.; Jardetzky, T. S.; Gorga, J. C.; Urban, R.
G.; Strominger, J. L.; Wiley: D. C. Nature 1994, 368, 215-221.
(24) Garboczi, D. N.; Ghosh, P.; Utz, U.; Fan, Q. R.; Biddison, W. E.;
Wiley: D. C. Nature 1996, 384, 134-141.
General Methods and Materials. All reactions were carried out
under an inert atmosphere with dry solvents under anhydrous conditions,
unless otherwise stated. CH2Cl2 and THF were distilled from calcium

Glycopeptide DeriVed from Type II Collagen
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7681
hydride and sodium-benzophenone, respectively. Pyridine was dried
over 4 Å molecular sieves; DMF was distilled and then sequentially
dried over two portions of 3 Å molecular sieves. TLC was performed
on silica gel 60 F254 (Merck) with detection by UV light and charring
with aqueous sulfuric acid or phosphomolybdic acid/ceric sulfate/
aqueous sulfuric acid. Flash column chromatography was performed
on silica gel (Matrex, 60 Å, 35-70 µm, Grace Amicon). Centrifugal
TLC was performed using rotors coated with silica gel 60 PF254
containing gypsum (Merck). The moving bands were visualized using
UV light. Organic solutions were dried over Na2SO4 before being
concentrated.
mg) was dissolved in 80% aqueous EtOH (15 mL), and the pH was
carefully adjusted to pH 7 with 30% aqueous Cs2CO3, concentrated at
reduced pressure, then diluted with EtOH (absolute), and concentrated
at high vacuum. The resulting powder was dissolved in DMF (6.5
mL) and cooled to 0 °C, after which benzyl bromide (301 µL, 2.5 mmol)
was added dropwise. After 12 h at room temperature the reaction
mixture was diluted with water and extracted with Et2O. The organic
phase was dried, and after concentration at reduced pressure the residue
was purified by centrifugal TLC fractionation (heptane-ethyl acetate,
1:1) to give 2 (1153 mg, 68% overall): [R]20D -6° (c 1.3, CHCl3); IR
1
(neat) 3360, 1743, 1688, 1534 cm-1; H NMR (400 MHz, CDCl3) δ
1
The H and 13C NMR spectra were recorded with a Bruker DRX-
5.61 (d, J ) 7.2 Hz, 1 H, NHR), 5.21 (ABd, 1 H, J ) 12.0 Hz,
PhCH2O), 5.16 (ABd, 1 H, J ) 12.1 Hz, PhCH2O), 4.90 (br s, 1 H,
NHꢀ), 4.49-4.41 (m, 1 H, HR), 4.41 (ABdd, 1 H, J ) 7.2, 10.6 Hz,
OCOCH2CH), 4.36 (ABdd, 1 H, J ) 7.3, 10.5 Hz, OCOCH2CH), 4.21
(t, J ) 7.1 Hz, 1 H, OCOCH2CH), 3.67 (br s, 1 H, Hδ), 3.26-3.19
(m, 1 H, Hꢀ), 3.01-2.90 (m, 1 H, Hꢀ), 2.08-2.00 (m, 1 H, Hâ), 1.83-
1.74 (m, 1 H, Hâ), 1.47-1.40 (m, 2 H, Hγ), 1.44 (s, 9 H, tBu); 13C
NMR (91 MHz, CDCl3) δ 172.2, 156.9, 156.1, 143.8, 143.7, 141.2,
135.2, 128.6, 128.5, 128.3, 127.6, 127.0, 125.0, 119.9, 79.7, 71.0, 67.2,
67.0, 53.7, 47.1, 46.5, 30.1, 28.9, 28.3; HRMS (FAB) calcd for
C33H39O7N2 575.2757 (M + H+), found 575.2756. Anal. Calcd for
C33H38O7N2: C, 69.0; H, 6.7; N, 4.9. Found: C, 68.9; H, 6.8; N, 4.9.
(5R)-Nr-(Fluoren-9-ylmethoxycarbonyl)-NE-(tert-butoxycarbonyl)-
5-O-(tert-butyldimethylsilyl)-5-hydroxy-L-lysine (3). 2,6-Lutidine (40
µL, 0.34 mmol) followed by tert-butyldimethylsilyl trifluoromethane-
sulfonate (58 µL, 0.26 mmol) was added to a solution of 2 (100 mg,
0.17 mmol) in CH2Cl2 (1 mL) at 0 °C, and the solution was stirred for
2 h. Et2O was added, and the solution was washed with water followed
by saturated aqueous NaCl. The organic phase was dried, concentrated,
and filtered through silica gel (heptane-ethyl acetate, 1:1) to give the
desired silyl ether (115 mg): 1H NMR (360 MHz, CDCl3) δ 5.47 (d,
1 H, J ) 6.3 Hz, NHR), 5.21 (ABd, 1 H, J ) 12.2 Hz, PhCH2O), 5.16
(ABd, 1 H, J ) 12.2 Hz, PhCH2O), 4.73 (br s, 1H, NHꢀ), 4.44-4.32
(m, 3 H, OCOCH2CH, HR), 4.22 (t, 1 H, J ) 7.0 Hz, OCOCH2CH),
3.74-3.68 (m, 1 H, Hδ), 3.19-3.02 (m, 2 H, Hꢀ), 2.03-1.93 (m, 1 H,
Hâ), 1.76-1.64 (m, 1 H, Hâ), 0.88 (s, 9 H, tBu), 0.07 and 0.03 (2 s,
each 3 H, SiCH3). The silyl ether from above (100 mg, 0.14 mmol) in
EtOAc (5 mL) was subjected to hydrogenolysis (1 atm) over 10% Pd-C
(40 mg) at room temperature for 1.5 h. The catalyst was then filtered
off (Hyflo-Supercel) and washed with EtOAc, EtOH, and MeOH. The
combined filtrates were concentrated, and centrifugal TLC fractionation
360, Bruker DRX-400, or Bruker ARX-500 spectrometer for solutions
in CDCl3 [residual CHCl3 (δH 7.27 ppm) or CDCl3 (δC 77.0 ppm) as
internal standard] or CD3OD [residual CD2HOD (δH 3.35 ppm) or
1
CD3OD (δC 49.0 ppm) as internal standard] at 300 K. The H NMR
spectrum of glycopeptide 15 was recorded with a Bruker ARX-500
spectrometer in a 9:1 mixture of H2O and D2O [H2O (δH 4.98 ppm) as
internal standard] at 278 K. First-order chemical shifts and coupling
constants were obtained from one-dimensional spectra, and proton
resonances were assigned from COSY,25 TOCSY,26 and ROESY27
experiments. Resonances for aromatic protons and resonances that
could not be assigned are not reported. IR spectra were recorded on
an ATI Mattson Genesis Series FTIR spectrometer. Optical rotations
were measured with a Perkin-Elmer 343 polarimeter. Ions for positive
fast atom bombardment mass spectra were produced by a beam of xenon
atoms (6 keV) from a matrix of glycerol and thioglycerol. In the amino
acid analyses, 5-hydroxynorvaline was not determined, and glutamine
was determined as glutamic acid.
Analytical normal phase HPLC was performed on a Kromasil silica
column (100 Å, 5 µm, 4.6 × 250 mm) with a flow rate of 2 mL/min
and detection at 254 nm. Preparative purifications were performed on
a Kromasil silica column (100 Å, 5 µm, 20 × 250 mm) with a flow
rate of 20 mL/min.
Compounds 1,12 4,13 5,12 6,12b 9,12b 10,12b and 1112b were prepared
as described in the cited references. Synthesis of glycopeptides 14,
16, and 17 has been described elsewhere.12
(5R)-Nr-(Fluoren-9-ylmethoxycarbonyl)-NE-(tert-butoxycarbonyl)-
5-hydroxy-L-lysine Benzyl Ester (2). Copper(II) carbonate (658 mg,
2.98 mmol) was added gradually at 85-90 °C to a solution of (5R)-
5-hydroxy-L-lysine dihydrochloride monohydrate (945 mg, 3.8 mmol)
in water (10 mL). The mixture was refluxed for 20 min and then
filtered (Hyflo-Supercel). The filter pad was washed with several small
portions of hot water until the filtrate was colorless. The volume of
the combined filtrate was adjusted to 25 mL with water, and then
NaHCO3 (662 mg, 7.9 mmol) was added at 20 °C. A solution of di-
tert-butyl dicarbonate (1.14 g, 5.2 mmol) in dioxane (10 mL) was added
over 1 h, and the reaction mixture was then stirred overnight at 20 °C.
The resulting suspension was filtered, and the filter cake was washed
with small portions of water. The combined filtrates were concentrated,
diluted with water/THF (10 mL, 1:1), and cooled to 5 °C. After 2 h
another portion of precipitate was filtered off, and the combined
precipitates were suspended in MeOH (15 mL) and stirred overnight.
Na+ Chelex 100 (25 mL) which had been converted to its H+ form9
was added together with water (20 mL). The resulting mixture was
gently swirled for 4 h and then filtered. The resin was washed with
several portions of 50% aqueous MeOH, and the combined filtrates
were concentrated at high vacuum to give the Nꢀ-tert-butoxycarbonyl-
protected derivative (764 mg, ∼78%) which was used in the next step
without purification. A solution of 9-fluorenylmethyl chloroformate
(588 mg, 2.28 mmol) in dioxane (3.5 mL) was added dropwise to a
solution of the Nꢀ-tert-butoxycarbonyl-protected derivative from above
(600 mg, 2.28 mmol) in dioxane/10% aqueous Na2CO3 (7 mL, 1:2) at
0 °C. The solution was stirred for 4 h at room temperature, cold water
(120 mL) was added, and the reaction mixture was carefully acidified
to pH 3 with cold 1 M aqueous KHSO4, then extracted with EtOAc,
dried, and concentrated at reduced pressure. The resulting oil (1050
(EtOAc-EtOH, 10:1 f 0:1) of the residue gave 3 (61 mg, 72%): [R]20
D
1
+3° (c 1, CHCl3); H NMR (400 MHz, CD3OD) δ 6.41 (t, 1 H, J )
5.7 Hz, NHꢀ), 4.34 (ABdd, 1 H, J ) 7.4, 10.4 Hz, OCOCH2CH), 4.30
(ABdd, 1 H, J ) 6.7, 10.5 Hz, OCOCH2CH), 4.20 (t, 1 H, J ) 6.9 Hz,
OCOCH2CH), 4.05 (br s, 1 H, HR), 3.78-3.73 (m, 1 H, Hδ), 3.06-
2.99 (m, 2 H, Hꢀ), 2.03-1.94 (m, 1 H, Hâ), 1.70-1.43 (m, 3 H,
Hâ′,γ,γ′), 0.88 (s, 9 H, tBu), 0.08 and 0.06 (2 s, each 3 H, SiCH3); 13
C
NMR (91 MHz, CD3OD) δ 177.7, 158.5, 158.4, 145.3, 145.2, 142.6,
128.8, 128.2, 126.3, 120.9, 80.0, 72.2, 67.9, 56.3, 48.4, 47.1, 32.5, 29.1,
28.8, 26.4, 18.9, -4.2, -4.4; HRMS (FAB) calcd for C32H46N2O7SiNa
621.2972 (M + Na+), found 621.2966.
(5R)-Nr-(Fluoren-9-ylmethoxycarbonyl)-NE-(tert-butoxycarbonyl)-
5-O-(6-O-(tert-butyldiphenylsilyl)-3,4-O-isopropylidene-â-D-galac-
topyranosyl)-5-hydroxy-L-lysine Benzyl Ester (7). A solution of
dimethyldioxirane28 in acetone (15.5 mL, ∼0.07 M, 0.78 mmol) was
added to 6-O-(tert-butyldiphenylsilyl)-3,4-O-isopropylidene-D-galactal13
(4; 274 mg, 0.646 mmol) in CH2Cl2 (7 mL) at 0 °C, and the solution
was protected from light and stirred at 0 °C for 60 min. Concentration
of the solution under reduced pressure yielded the corresponding R-1,2-
anhydrosugar. A solution of 2 (743 mg, 1.29 mmol) in THF (8 mL)
containing crushed molecular sieves (AW-300, 500 mg) was added to
the R-1,2-anhydrosugar, and the mixture was stirred for 30 min at room
temperature and then cooled to -50 °C. Zinc chloride (970 µL, 1.0
M in Et2O, 0.970 mmol) was added, and the reaction mixture was
allowed to attain room temperature over 19 h. The mixture was then
diluted with EtOAc, filtered (Hyflo-Supercel), and washed with water.
The aqueous phase was twice extracted with EtOAc, and the combined
(25) Derome, A. E.; Williamson, M. P. J. Magn. Reson. 1990, 88, 177-
185.
(26) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355-360.
(27) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 63, 207-213.
(28) Adam, W.; Bialas, J.; Hadjiarapoglou, L. Chem. Ber. 1991, 124,
2377.

7682 J. Am. Chem. Soc., Vol. 120, No. 31, 1998
Broddefalk et al.
organic phases were dried and concentrated. Flash column chroma-
tography (toluene-MeCN, 6:1 f 4:1) followed by purification by
normal-phase HPLC (linear gradient 0 f 20% tert-butyl methyl ether
in CH2Cl2 during 160 min) gave 7 (242 mg, 37%), the corresponding
R anomer (30 mg, 5%), and unreacted acceptor 2 (421 mg, 0.733 mmol).
Data for Compound 7: [R]20D +8° (c 1.0, CHCl3); 1H NMR (500 MHz,
CDCl3) δ 5.59 (d, J ) 7.4 Hz, 1 H, NHR), 5.36 (br s, 1 H, NHꢀ), 5.20
(ABd, 1 H, J ) 12.1 Hz, PhCH2O), 5.16 (ABd, 1 H, J ) 12.2 Hz,
PhCH2O), 4.44-4.38 (m, 1 H, HR), 4.42 (ABdd, 1 H, J ) 7.0, 10.6
Hz, OCOCH2CH), 4.35 (ABdd, 1 H, J ) 7.2, 10.7 Hz, OCOCH2CH),
4.23 (br d, 1 H, J ) 4.1 Hz, H-4), 4.19 (t, J ) 6.9 Hz, 1 H,
OCOCH2CH), 4.15 (d, 1 H, J ) 8.1 Hz, H-1), 4.02 (br t, 1 H, J ) 6.3
Hz, H-3), 3.95-3.91 (m, 2 H, H-6), 3.90-3.86 (m, 1 H, H-5), 3.63 (br
s, 1 H, Hδ), 3.51 (t, 1 H, J ) 7.7 Hz, H-2), 3.29-3.23 (m, 1 H, Hꢀ),
3.18 (br s, 1 H, OH), 3.10 (dt, 1 H, J ) 5.7, 14.4 Hz, Hꢀ), 2.01-1.93
(m, 1 H, Hâ), 1.92-1.84 (m, 1 H, Hâ), 1.66-1.58 (m, 2 H, Hγ), 1.49
bance of bromophenol blue at 600 nm, and the peptide-resin was
automatically washed with DMF after 1 h or when monitoring revealed
the coupling to be complete. NR-Fmoc deprotection of the peptide resin
was performed by a flow of 20% piperidine in DMF (2 mL/min)
through the column for 12.5-27.5 min, and was monitored32 using the
absorbance of the dibenzofulvene-piperidine adduct at 350 nm. After
completion of the NR-Fmoc deprotection, the peptide-resin was again
washed automatically with DMF. The glycosylated amino acid 8 (72
µmol) was activated separately in DMF (1.0 mL) at room temperature
during 30 min by addition of 1,3-diisopropylcarbodiimide (72 µmol)
and 1-hydroxy-7-azabenzotriazole33 (HOAt, 0.216 mmol). Compound
8 was then coupled manually to the peptide-resin which had been
removed from the synthesizer. The coupling of 8 was performed in a
mechanically agitated reactor during 24 h, and it was monitored by
bromophenol blue as described above. After coupling of 8 the
glycopeptide resin was reinserted in the synthesizer, and coupling of
the remaining amino acids was performed as outlined above. Peptide
13 was synthesized essentially in the same manner.
After completion of the synthesis, the resins carrying protected
peptide 13 and glycopeptide 15 were washed with CH2Cl2 and dried
under vacuum. A portion of each (glyco)peptide-resin was cleaved
(cf. details given below for 13 and 15), the amino acid side chains
were deprotected, and acid-labile carbohydrate protective groups were
removed by treatment with trifluoroacetic acid/water/thioanisole/
ethanedithiol [87.5:5:5:2.5, ∼20 mL/200 mg of (glyco)peptide resin]
for 2 h followed by filtration. Acetic acid (5 mL) was added to the
filtrate, the solution was concentrated, and acetic acid (2 × 5 mL) was
added again followed by concentration after each addition. The residue
was triturated with diethyl ether (15 mL) which gave a solid, crude
glycopeptide which was dissolved in a mixture of acetic acid and water
(10 mL) and freeze-dried. Purification by preparative HPLC gave pure
13 and 15.
and 1.33 (2 s, each 3 H, CH3), 1.28 and 1.04 (2 s, each 9 H, tBu); 13
C
NMR (126 MHz, CDCl3) δ 172.0, 156.2, 156.0, 143.8, 143.7, 141.2,
135.6, 135.5, 135.1, 133.2, 133.0, 129.8, 128.6, 128.5, 128.4, 127.8,
127.7, 127.1, 125.1, 119.9, 110.0, 102.8, 80.2, 79.0, 78.9, 73.6, 73.1,
67.4, 67.1, 62.5, 53.7, 47.2, 44.2, 28.4, 28.2, 28.1, 26.8, 26.3, 19.2;
HRMS (FAB) calcd for C58H70N2O12SiNa 1037.4596 (M + Na+), found
1037.4608. Anal. Calcd for C58H70O12N2Si: C, 68.6; H, 7.0; N, 2.8.
Found: C, 68.7; H, 7.0; N, 2.9.
(5R)-Nr-(Fluoren-9-ylmethoxycarbonyl)-NE-(tert-butoxycarbonyl)-
5-O-(6-O-(tert-butyldiphenylsilyl)-3,4-O-isopropylidene-â-D-galac-
topyranosyl)-5-hydroxy-L-lysine (8). A solution of 7 (198 mg, 0.195
mmol) in EtOAc (12 mL) was treated with 10% Pd-C (198 mg) under
hydrogen (1 atm) at room temperature for 4 h. The catalyst was then
filtered off (Hyflo-Supercel) and washed with EtOAc and EtOH. The
combined filtrates were concentrated, and flash column chromatography
(toluene-EtOH, 20:1 f 3:1) of the residue gave 8 (154 mg, 85%):
1
[R]20 +10° (c 0.8, CHCl3); H NMR (400 MHz, CD3OD) δ 4.37 (d,
Peptide 13 and glycopeptide 15 were analyzed on a Kromasil C-8
column (100 Å, 5 µm, 4.6 × 250 mm) using a linear gradient of 0 f
100% of B in A over 60 min with a flow rate of 1.5 mL/min and
detection at 214 nm (solvent systems: A, 0.1% aqueous trifluoroacetic
acid; B, 0.1% trifluoroacetic acid in CH3CN). Purification of crude
13 and 15 was performed on a Kromasil C-8 column (100 Å, 5 µm, 20
× 250 mm) using the same eluant and a flow rate of 11 mL/min.
The peptide content of the purified glycopeptide, as determined by
amino acid analysis, has been taken into account in calculating the
final yields for the glycopeptides. For example, 54 mg of 15 with a
peptide content of 72% was obtained, and the yield was therefore based
on 38.9 mg (54 mg × 0.72).
D
1 H, J ) 7.4 Hz, OCOCH2CH), 4.36 (d, 1 H, J ) 6.8 Hz, OCOCH2-
CH), 4.32-4.30 (m, 1 H, H-4), 4.27 (d, 1 H, J ) 8.2 Hz, H-1), 4.23
(br t, 1 H, J ) 7.0 Hz, OCOCH2CH), 4.15-4.11 (m, 1 H, HR), 4.02
(br t, 2 H, J ) 6.2 Hz, H-3, H-5), 3.89 (d, 2 H, J ) 6.9 Hz, H-6),
3.71-3.66 (m, 1 H, Hδ), 3.42 (dt, 1 H, J ) 7.7 Hz, H-2), 3.21 (ABdd,
1 H, J ) 3.9, 13.6 Hz, Hꢀ), 3.08 (ABdd, 1 H, J ) 7.1, 13.7 Hz, Hꢀ),
2.10-2.01 (m, 1 H, Hâ), 1.80-1.71 (m, 1 H, Hâ), 1.66-1.60 (m, 2
H, Hγ), 1.48 and 1.33 (2 s, each 3 H, CH3), 1.25 and 1.02 (2 s, each
9 H, tBu); 13C NMR (126 MHz, CD3OD) δ 178.0, 158.6, 158.2, 145.4,
145.2, 142.6, 136.8, 136.7, 134.4, 134.2, 131.0, 128.9, 128.8, 128.2,
126.3, 120.9, 110.8, 103.7, 80.7, 80.6, 80.0, 74.7, 74.6, 74.3, 67.9, 63.8,
56.4, 45.5, 30.2, 28.8 28.5, 27.3, 26.7, 20.0; HRMS (FAB) calcd for
C51H64N2O12SiNa 947.4126 (M + Na+), found 947.4142.
General Procedure for Solid-Phase Peptide Synthesis. Peptide
13 and glycopeptide 15 were synthesized in a custom-made, fully
automatic continuous flow peptide synthesizer constructed essentially
as described.29 A resin consisting of a cross-linked polystyrene
backbone grafted with poly(ethylene glycol) chains was used for the
syntheses. The resin carried the C-terminal lysine on a p-hydroxy-
methylphenoxy linker (TentaGel S PHB, Rapp Polymere, Germany).
NR-Fmoc-amino acids (Bachem, Switzerland) with the following
protective groups were used: triphenylmethyl (Trt) for glutamine; tert-
butyl for glutamic acid and hydroxyproline; and tert-butoxycarbonyl
(Boc) for lysine. DMF was distilled before being used.
Glycyl-L-glutam-1-yl-trans-4-hydroxy-L-prolylglycyl-L-isoleucyl-
L-alanylglycyl-L-phenylalanyl-5-hydroxy-L-lysylglycyl-L-glutam-1-
yl-L-glutaminylglycyl-L-prolyl-L-lysine (13). Synthesis, cleavage of
the resin-bound peptide (133 mg, 25 µmol) with simultaneous depro-
tection, and then purification by reversed-phase HPLC (linear gradient
0 f 100% B in A during 60 min), according to the general procedure,
gave 13 (10 mg, 74% peptide content, 20% overall yield): MS (FAB)
calcd 1504 (M + H+), found 1503; amino acid analysis, Ala 0.98 (1),
Glu 2.96 (3), Gly 5.01 (5), Hyl 1.00 (1), Hyp 1.02 (1), Ile 1.01 (1),
Lys 1.01 (1), Phe 0.99 (1), Pro 1.01 (1).
Glycyl-L-glutam-1-yl-trans-4-hydroxy-L-prolylglycyl-L-isoleucyl-
L-alanylglycyl-L-phenylalanyl-5-O-(â-D-galactopyranosyl)-5-hydroxy-
L-lysylglycyl-L-glutam-1-yl-L-glutaminylglycyl-L-prolyl-L-lysine (15).
Synthesis, cleavage of the resin-bound glycopeptide (368 mg, 60 µmol)
with simultaneous deprotection, and then purification by reversed-phase
HPLC (linear gradient 0 f 100% B in A during 60 min), according to
the general procedure, gave 15 (54 mg, 72% peptide content, 39%
overall yield): 1H NMR data, see Table 1; MS (FAB) calcd 1666 (M
+ H+), found 1666; amino acid analysis, Ala 1.00 (1), Glu 2.93 (3),
Gly 5.04 (5), Hyl 0.97 (1), Hyp 1.02 (1), Ile 1.02 (1), Lys 1.00 (1),
Phe 1.01 (1), Pro 1.03 (1).
In the synthesis of glycopeptide 15, 60 µmol of resin was used in
the peptide synthesizer. The NR-Fmoc-amino acids were activated as
1-benzotriazolyl esters.30 These were prepared in situ by reaction of
the appropriate NR-Fmoc-amino acid (0.24 mmol), 1-hydroxybenzo-
triazole (HOBt) (0.36 mmol), and 1,3-diisopropylcarbodiimide (0.234
mmol) in DMF (1.3 mL). After 45 min bromophenol blue (51 nmol,
0.3 mL of a 0.15 mM solution in DMF) was added to the 1-benzot-
riazolyl ester solution which was then recirculated through the column
containing the resin. The acylation was monitored31 using the absor-
Determination of T Cell Hybridoma Response. The response of
each T cell hybridoma, i.e., IL-2 secreted on incubation of the
hybridoma with antigen-presenting spleen cells and increasing con-
(29) Cameron, L. R.; Holder, J. L.; Meldal, M.; Sheppard, R. C. J. Chem.
Soc., Perkin Trans. 1 1988, 2895-2901.
(30) Ko¨nig, W.; Geiger, R. Chem. Ber. 1970, 103, 788-798.
(31) Flegel, M.; Sheppard, R. C. J. Chem. Soc., Chem. Commun. 1990,
536-538.
(32) Dryland, A.; Sheppard, R. C. J. Chem. Soc., Perkin Trans. 1 1986,
125-137.
(33) Carpino, L. A. J. Am. Chem. Soc. 1993, 115, 4397-4398.

Glycopeptide DeriVed from Type II Collagen
J. Am. Chem. Soc., Vol. 120, No. 31, 1998 7683
centrations of antigen [type II collagen or (glyco)peptides 12-17], was
determined in a standard assay using the T cell clone CTLL.21 In brief,
5 × 104 T cell hybridomas were cocultured with 5 × 105 syngeneic,
irradiated (1500 rad) spleen cells and antigen in a volume of 200 µL
in flat-bottom microtiter plate wells. After 24 h, 100 µL aliquots of
the supernatants were removed and frozen to kill any transferred T
cell hybridomas. To the thawed supernatant, 104 IL-2 sensitive CTLL
T cells were added. The CTLL cultures were incubated for 24 h, after
Acknowledgment. This work was funded by grants from
the Swedish Natural Science Research Council, the Swedish
Medical Research Council, the Craaford Foundation, the Anna
Greta Craaford Foundation for Rheumatological Research, the
Nanna Svartz Foundation, the King Gustaf V:s 80-year Founda-
tion, the Kock and O¨ sterlund Foundations, and the Swedish
Association against Rheumatism.
3
which they were pulsed with 1 µCi of H-TdR for an additional 15-
Supporting Information Available: Spectra of 3 and 8 (4
pages, print/PDF). See any current masthead page for ordering
information and Web access instructions.
18 h. The cells were harvested on glassfiber sheets in a Filtermate
TM cell harvester (Packard Instruments, Meriden, CT), and the amount
of radioactivity was determined in a matrix 96 Direct Beta Counter
(Packard). All experiments were performed in duplicate.
JA980489K